REAL TIME RECOGNITION OF THE CORRELATION PEAKS OF MULTI CHANNEL OPTICAL CORRELATOR USING NEURAL NETWORK
基于人工神经网络的多路光学相关器相关峰的实时识别

A method using neural network to recognize targets according to the output of the multichannel optical correlator was proposed. The position of the correlation peaks was found through statistical rules. Not only the self correlation peaks but also the cross correlation peaks were used to be the features of the targets and the complex feature extraction was realized through neural network. The proposed method can effectively exclude the effects of the target background interference and the fluctuation of the device performance so as to realize a real time target recognition.